Recently, a number of nucleic acid-based strategies have been developed to modulate a variety of cellular functions (Opalinska and Gewirtz, 2002). Classes of oligonucleotides such as aptamers, transcription factor-binding decoy oligonucleotides, ribozymes, triplex-forming oligonucleotides, immunostimulatory CpG motifs, antisense oligonucleotides (including peptide nucleic acids), small interfering RNAs, and microRNAs have drawn much interest as research tools, owing to their highly specific mode of action. These oligomeric nucleic acids have considerable potential as therapeutics, too. Such therapeutics face several obstacles, however, including the instability of free nucleic acids and the safe, efficient and targeted cellular delivery of these macromolecules (Dykxhoorn and Lieberman, 2005).
A focus of many nucleic acid-based therapeutic strategies is the phenomenon of RNA interference (RNAi), whereby long, double-stranded RNA (dsRNA) in a cell leads to sequence-specific degradation of homologous (complementary or partially complementary) gene transcripts. More particularly, the long dsRNA molecules are processed into smaller RNAs by an endogenous ribonuclease called “Dicer” (Grishok et al., 2000; Zamore et al., 2000). The smaller RNAs are known as “short interfering RNA” (siRNA) when they derive from exogenous sources and as “microRNA” (miRNA) when they are produced from RNA-coding genes in the cell's own genome. These two classes of small (typically, 21- to 23-nucleotide) regulatory RNAs also differ in that miRNAs show only partial complementarity to messenger RNA (mRNA) targets.
The short regulatory RNAs bind to the so-called “RNA-induced silencing complex” (RISC), which has a helicase activity and an endonuclease activity. The helicase activity unwinds the two strands of RNA molecules, allowing the antisense strand to bind to the targeted RNA molecule (Zamore et al., 2000; Zamore, 2002; Vickers et al., 2003). The endonuclease activity hydrolyzes the target RNA at the site where the antisense strand is bound.
In RNAi, therefore, a single-stranded RNA molecule (ssRNA) binds to the target RNA molecule by Watson-Crick base-pairing rules and recruits a ribonuclease that degrades the target RNA. By contrast, antisense suppression of gene expression entails the binding of ssRNA to mRNA, blocking translation without catalyzing the degradation of the mRNA.
As a class, regulatory RNAs have a half-life of less than an hour in human plasma (Layzer et al., 2004) and are rapidly excreted by the kidneys. Consequently, several groups have attempted to prepare regulatory RNAs, including siRNAs, that are nuclease-resistant. Examples of such efforts include chemically modifying the nucleotides (e.g., 2′-F, 2′-OMe, Locked Nucleic Acids; LNA) or the phosphodiester backbone, e.g., phosphorothioate linkages (Chiu and Rana 2003; Choung et al., 2006; Czauderna et al. 2003; Elmén et al., 2005; Layzer et al., 2004; Morrissey et al., 2005). Also, to minimize the time siRNAs or other regulatory RNAs spend in circulation, practitioners have conjugated the RNA molecules to proteins and antibodies to target desired mammalian cells. In further efforts to address the concerns of low stability and rapid renal excretion, practitioners have developed vehicles for delivering regulatory RNAs. Polyplexes (formed by self assembly of nucleic acids with polycations), lipopolyplexes (formed by initial condensation of the nucleic acid with polycations, followed by addition of cationic lipids), liposomes, and synthetic nanoparticles are being explored, too.
These approaches also face numerous obstacles, such as (a) rapid clearance of the carrier proteins from the serum through renal excretion, (b) limited number of regulatory RNA molecules that can be conjugated to each carrier protein, (c) difficulty in intracellular dissociation of intact, regulatory RNAs from the carrier protein, (d) rapid clearance due to polyplexes binding serum proteins which can act as opsonins (Dash et al., 1999), and (e) instability of liposomes in vivo, causing release of nucleic acids into the serum and potential non-specific transformation.
Viral vectors also have been developed to produce regulatory RNAs endogenously. See, e.g., Devroe and Silver, 2004. These viral vectors pose serious safety concerns, however. Illustrative problems include recombination with wild-type viruses, insertional and oncogenic potential, virus-induced immunosuppression, limited capacity of the viral vectors to carry large segments of DNA, reversion to virulence of attenuated viruses, difficulties in manufacture and distribution, low stability, and adverse reactions (Hacein-Bey-Abina et al., 2003; Kootstra and Verma, 2003; Raper et al., 2003; Verma and Weitzman, 2005; Check, 2005).
Plasmid-based systems also have been developed for recombinant, in situ expression of a regulatory RNA, such as an siRNA or a larger (˜70 nt) precursor, a short hairpin RNA (shRNA). An shRNA contains sense and antisense sequences from a target gene that are connected by a hairpin loop. See, e.g., Paddison et al., 2002. shRNAs can be expressed from a pol-III-type promoter or, in the context of a miRNA, by pol II promoters.
As described in international application WO 03/033519, plasmids that code for an shRNA, siRNA, or other regulatory RNA can be transformed into a parent bacterial strain that produces intact minicells, by virtue of a mutation that causes asymmetric cell division. Such transformation yields recombinant bacteria in which the plasmid replicates intracellularly, introducing large numbers of plasmids in the bacterial cytoplasm. During the asymmetric division, some of the plasmids segregate into the minicell cytoplasm, resulting in recombinant minicells. The minicells then can deliver the plasmid DNA into a mammalian cell, where the plasmid DNA migrates to the cell nucleus. In the nucleus the plasmid DNA expresses the shRNA or other regulatory RNA, as the case may be, and the resultant nucleic acid then migrates to the cytoplasm, where it can effect RNAi or gene suppression, depending on the nature of the involved regulatory RNA.
Because such approaches require host machinery, however, delivering therapeutically effective amounts of nucleic acid via expression-based systems involves complex and protracted processes, which limits their effectiveness. Accordingly, a more efficacious methodology is needed for delivering functional nucleic acids, such as regulatory RNAs, to target cells.